Febraury 2, 2021
The purpose of this capstone project was to build a Shiny App for text prediction, based on the SwiftKey database.
In the first part of this project (Milestone Report) I reported a simple data and exploratory analysis of the database in object and with this I built a Corpora for text prediction.
The app was built using R Studio and Shiny Apps, and is based on a sample of 10%, since it has to run also on low-powered devices.
The app, deployed on ShinyApps, takes as input one or more words and then returns the most frequent word using n-grams which are a sequence of items, in this case words.
In particular, the three databases on which the app is based are aggregated bi/tri/quad-grams of words.
The image below shows the app interface, with just an input box, and a return field to display the predicted word.
Coursera Data Science Specialization
Shinyapps
Swiftkey